CFD MODELING OF LH 2 DISPERSION USING THE

CFD MODELING OF LH 2 DISPERSION USING THE ADREA-HF CODE Giannissi, S. G. 1, 2, Venetsanos, A. G. 1, Bartzis 3, J. G. , Markatos 2, N. , Willoughby, D. B. 4 and Royle, M. 4 Environmental Research Laboratory, National Centre for Scientific Research Demokritos, 15310 Aghia Paraskevi, Attikis, Greece, email: sgiannissi@ipta. demokritos. gr, venets@ipta. demokritos. gr 1 National Technical University of Athens, School of Chemical Engineering, Department of Process Analysis and Plant Design, Heroon Polytechniou 9, 15780 Zografou, Greece, email: n. markatos. @ntua. gr 2 Department of Energy and Resources Management Engineering, University of West Macedonia, Kozani, Greece, email: bartzis@uowm. gr 3 Health and Safety Laboratory, Buxton, Derbyshire, SK 17 9 JN, United Kingdom, email: Deborah. Willoughby@hsl. gov. uk, Mark. Royle@hsl. gov. uk 4 4 TH International Conference on Hydrogen Safety (ICHS 2011) September 12 -14, 2011, San Francisco, USA organized by IA Hy. Safe 1

OUTLINE • Objectives • HSL (Health and Safety Laboratory) Experiments • Test 1 (test chosen for simulation) o Test 1 -Humidity Effect • Modeling Strategy o Physics o Numerics • Results • Conclusions 4 TH International Conference on Hydrogen Safety (ICHS 2011) September 12 -14, 2011, San Francisco, USA organized by IA Hy. Safe 2/19

OBJECTIVES • Validation of the CFD code, ADREA-HF, for its performance in simulation of cryogenic releases. Test 1 of the HSL experiments (LH 2 release experiments) is chosen for simulation. • Examination of the humidity effect on the hydrogen vapor dispersion. Water vapor liquefaction and solidification due to the cold, hydrogen cloud (20 K) Heat liberation (latent heat of liquefaction and solidification) 4 TH International Conference on Hydrogen Safety (ICHS 2011) September 12 -14, 2011, San Francisco, USA H 2 vapor cloud more buoyant organized by IA Hy. Safe 3/19

HSL (HEALTH AND SAFETY LABORATORY) EXPERIMENTS 1 • 4 LH 2 release tests with spill rate 60 lt/min Test Release height Release direction Spill duration (sec) 1 3. 36 horizontal 248 2 100 vertically downwards 561 3 860 horizontal 305 4 100 vertically downwards 215 (mm above ground) Photograph taken from HSL 1 Willoughby, D. B. , Royle, M. , Experimental Releases of Liquid Hydrogen, 4 th International Conference on Hydrogen Safety, San Francisco, California-USA, ICHS , Paper 1 A 3, 2011 4 TH International Conference on Hydrogen Safety (ICHS 2011) September 12 -14, 2011, San Francisco, USA organized by IA Hy. Safe 4/19

TEST 1 • Release and weather conditions source diameter (mm) 26. 6 source temperature (K) 20 release rate (kg/sec) 0. 07 release duration (sec) 248 wind speed (m/s)-@2. 5 2. 675 wind direction-@ 2. 5 m 291. 02 average ambient temperature (K) 283. 56 relative humidity (%) 68 Photographs taken from HSL 4 TH International Conference on Hydrogen Safety (ICHS 2011) September 12 -14, 2011, San Francisco, USA organized by IA Hy. Safe 5/19

TEST 1 • Release and weather conditions source diameter (mm) 26. 6 source temperature (K) 20 release rate (kg/sec) 0. 07 release duration (sec) 248 wind speed (m/s)-@2. 5 2. 675 wind direction-@ 2. 5 m 291. 02 average ambient temperature (K) 283. 56 relative humidity (%) 68 Site layout (not drawn to scale) 4 TH International Conference on Hydrogen Safety (ICHS 2011) September 12 -14, 2011, San Francisco, USA organized by IA Hy. Safe 6/19

MODELING STRATEGY Physics (1/3) • Multi-phase multi component RANS CFD calculation using ADREAHF CFD code. • 3 -D transient, fully compressible conservation equations for mixture mass, mixture momentum, mixture enthalpy, hydrogen mass fraction and water mass fraction (when ambient humidity was taken into account). • Phase distribution: Non vapor phase (liquid+solid) of component-I appears when the mixture temperature falls below the mixture dew temperature, which is calculated using the Raoult’s law for ideal gases. The solid phase of component-I appears when the mixture temperature drops below the freezing point. • Standard k-ε with buoyancy effect term. • One dimensional, transient energy (temperature) equation inside the ground. The ground has the concrete’s properties. 4 TH International Conference on Hydrogen Safety (ICHS 2011) September 12 -14, 2011, San Francisco, USA organized by IA Hy. Safe 7/19

MODELING STRATEGY Physics (2/3) • In presence of solid H 2 O (ice), mixture dynamic viscosity is calculated using 2 different approaches: o Ice viscosity function of temperature • The liquid H 2 O viscosity correlation used below the FP o Constant ice viscosity • Equal to the water viscosity at freezing point 4 TH International Conference on Hydrogen Safety (ICHS 2011) September 12 -14, 2011, San Francisco, USA organized by IA Hy. Safe 8/19

MODELING STRATEGY Physics (3/3) Initial conditions: • To obtain the initial conditions of wind speed, ambient temperature and turbulence the procedure that followed consists of two steps: 1. One dimensional (in the z-direction) problem was solved to obtain the wind profile according to the experimental data. Neutral atmospheric conditions were assumed. 2. Three dimensional, steady problem was solved with initial conditions the ones calculated by the previous step (the wind direction was in line with the release). • The transient problem with hydrogen release was solved using as initial conditions the ones derived by the second step. In the case with humidity, additional initial condition for the water vapor mass fraction (5. 34∙ 10 -3) was used in the whole domain, calculated by the experiment’s relative humidity. Boundary conditions: • Inlet: The values of all variables were the same as the initial conditions. • Source: The source was modeled as two phase jet. The void fraction of the vapor phase is calculated by assuming isenthalpic expansion from 2 bars (inside the tanker) to 1. 2 bars (after the valve is open) and is equal to 71. 34%. Temperature, pressure and horizontal velocity were set equal to 21 K, 1. 2 bars and 6. 02 m/s respectively. 4 TH International Conference on Hydrogen Safety (ICHS 2011) September 12 -14, 2011, San Francisco, USA organized by IA Hy. Safe 9/19

MODELING STRATEGY • Numerics First order fully implicit scheme for time integration. First order upwind scheme for discretization of the convective terms ILU(0) preconditioned Bi. CGStab solver for the algebraic systems (parallel) Initial time step 10 -4 Courant number restriction (CFL<2) • • Figure from Edes (GUI of ADREA-Hf code) Computational domain (m) Grid characteristics (m) grid dimension x y 70 20 dx (min-max) dy (min-max) dz (min-max) 66 x 23 100188 0. 1 -6. 488 0. 1 -3. 867 0. 2 -2. 270 z 80 total number of cells 4 TH International Conference on Hydrogen Safety (ICHS 2011) September 12 -14, 2011, San Francisco, USA organized by IA Hy. Safe 10/19

RESULTS (1/6) Hydrogen concentration history at locations downwind the release point 4 TH International Conference on Hydrogen Safety (ICHS 2011) September 12 -14, 2011, San Francisco, USA organized by IA Hy. Safe 11/19

RESULTS (2/6) Hydrogen concentration history at different heights 4 TH International Conference on Hydrogen Safety (ICHS 2011) September 12 -14, 2011, San Francisco, USA organized by IA Hy. Safe 12/19

RESULTS (3/6) Duration 12 sec no humidity 4 TH International Conference on Hydrogen Safety (ICHS 2011) September 12 -14, 2011, San Francisco, USA humidity organized by IA Hy. Safe 13/19

RESULTS (4/6) t = 20 sec no humidity 4 TH International Conference on Hydrogen Safety (ICHS 2011) September 12 -14, 2011, San Francisco, USA humidity organized by IA Hy. Safe 14/19

RESULTS (5/6) H 2 vapor volume fraction contours H 2 O non vapor mass fraction contours t = 20 sec 4 TH International Conference on Hydrogen Safety (ICHS 2011) September 12 -14, 2011, San Francisco, USA organized by IA Hy. Safe 15/19

RESULTS (6/6) Temperature contours H 2 O non vapor mass fraction contours t = 20 sec 4 TH International Conference on Hydrogen Safety (ICHS 2011) September 12 -14, 2011, San Francisco, USA organized by IA Hy. Safe 16/19

CONCLUSIONS (1/2) Ø Multi-phase, multi component CFD calculations have been performed with ADREA-HF code to simulate HSL test-1 LH 2 release. The working fluid was assumed to be composed of dry air (gas), water (vapor/liquid/solid) and h 2 (vapor/liquid) Ø Predicted concentration histories with humidity are in better agreement with the experiment compared to the case without humidity. Ø It has been verified that the H 2 -humid air cloud becomes more buoyant than when neglecting the humidity, due to the heat liberation by the water vapor condensation/solidification. Ø Predictions with humidity were found sensitive to the way mixture molecular viscosity is modeled in case of presence of solids (ice). The assumption that ice viscosity follows the liquid viscosity formula below the freezing point gave good results. 4 TH International Conference on Hydrogen Safety (ICHS 2011) September 12 -14, 2011, San Francisco, USA organized by IA Hy. Safe 17/19

CONCLUSIONS (2/2) Ø Predictions show that including humidity reduces horizontal distance to LFL cloud by 40% (almost 10 m) and increase the height to LFL cloud (almost 1 m) in the present case. Ø Further work on the humidity effects is necessary to support present findings 4 TH International Conference on Hydrogen Safety (ICHS 2011) September 12 -14, 2011, San Francisco, USA organized by IA Hy. Safe 18/19

THANK YOU FOR YOUR ATTENTION ANY QUESTIONS 4 TH International Conference on Hydrogen Safety (ICHS 2011) September 12 -14, 2011, San Francisco, USA organized by IA Hy. Safe